Research on the Fusion and Analysis Method for the Multi-Source Heterogeneous Elevator Risk Data Based on Deep Learning

With the acceleration of urbanization, elevators, as essential vertical transportation tools, are increasingly the focus of public safety concerns. Due to improper operation, equipment defects, safety component failures, and other factors, elevator accidents frequently occur, posing a severe threat to people’s lives and property. This paper addresses the urgent needs for smart supervision and "on-demand maintenance" reform of elevators in China by researching a deep learning-based method for fusing and analyzing multi-source heterogeneous risk data of elevators. By integrating data from the Internet of Things, intelligent cameras, basic databases, and data exchange systems, this study constructs a risk early warning model aimed at achieving precise supervision and decision-making assistance. The research results indicate that the model can effectively identify potential risks in elevator operation, providing scientific decision support for regulatory authorities and promoting technological innovation and industry upgrading in the elevator sector.